Reliablesoft Academy How to Generate Images with Claude
0 / 8
AI Builders Lab > Resources > How to Generate Images with Claude

How to generate images with Claude. No subscriptions.

Every prompt from the video, in the order it was used. You connect your own Google (Nano Banana) and OpenAI (GPT Image) API keys to the Code tab of the Claude desktop app, then create photos, edit them, make a poster with text and build a small website with its own hero image, just by chatting. You only pay for the images you make.

8 steps 6 prompts 59 cents for the whole demo no image subscription no code to write

Get the full system for building with AI, including the workflows, the prompt library, and everything that would not fit inside one video.

AI Builders Lab →
  1. Get two API keys. One from Google for Nano Banana and one from OpenAI for GPT Image. Steps 1 and 2 show where to click.
  2. Save them in a Claude environment. The keys stay encrypted on your computer and never show up in your chats. Step 3.
  3. Run the setup prompt once. Claude builds a small image tool in your folder and leaves itself instructions on how to use it. Step 4.
  4. Then just ask. Describe an image, ask for a change, or ask for GPT Image when you need text. Every chat in that environment can make images.
You need a paid Claude plan (Pro or above), because this uses the Code tab of the Claude desktop app. Don't let the name put you off: you won't write any code, you just chat with Claude the way you normally do.

Claude can already design graphics like social posts, posters, infographics and slides. It builds them in code (HTML), so the text is always sharp and spelled correctly. What code can't make is a photo. For that you need an image model, and that's what these steps connect.
Step 1

Get A Google API Key

This key gives Claude access to Google's Nano Banana image models. There are no prompts in this step.

Create the key // aistudio.google.com
  • Go to aistudio.google.com and sign in with your Google account.
  • Click Get API key, then Create API key, and give it a name you'll recognize later. In the video it's called Claude images.
Turn on billing // this trips a lot of people up
  • A brand new key won't make images straight away, because Google's image models aren't on the free tier.
  • Next to your key, click Set up billing. Google bills through a Google Cloud billing account, so pick one you already have or add a new one with your card.
  • Open Spend in the menu on the left and set a monthly spend cap, so you can never spend more than you planned.
Then

Copy your key and keep it somewhere safe for a minute. Treat it like a password, because anyone who has it can spend your money.

Step 2

Get An OpenAI API Key

This key gives Claude access to OpenAI's GPT Image. Your ChatGPT login works, but a ChatGPT Plus subscription doesn't include the API, so this is billed separately. There are no prompts in this step.

Add credit // platform.openai.com
  • Go to platform.openai.com and sign in.
  • Open Settings, then Billing, and add $5 of credit. At medium quality, that's enough for hundreds of images.
  • Depending on your account, OpenAI may ask you to verify your organization before you can use the image models. It's a quick ID check.
Create the key
  • Go to API keys and click Create new secret key.
  • Give it a name, pick the default project, and choose when the key should expire.
  • OpenAI only shows you the key once, so copy it right away.
Step 3

Connect The Keys To Claude

An environment is a saved setup that tells Claude which folder to work in and what settings to use. You'll put your two keys in one, so every chat you start there can use them.

Make the folder
  • Make a new folder on your desktop called Image Studio. Every image Claude makes is saved here, so you'll always know where to find them.
Create the environment // claude desktop app → code
  • Open the Claude desktop app and click Code at the top.
  • Under the message box, click the environment menu and choose New environment.
  • Choose Local, name it Image Studio, and add the folder you just made.
  • Click Environment variables and paste the two lines below, one on each line. Replace the placeholders with your own keys.
  • Click Done, then Create environment.
Environment variables // replace the placeholders with your keys
GEMINI_API_KEY=your-google-key
OPENAI_API_KEY=your-openai-key
What you should see

Image Studio in the environment menu. Claude keeps the keys encrypted on your computer, so they never show up in your chats.

Step 4

Run The Setup Prompt

This prompt asks Claude to build a small image tool inside your folder, and to leave itself instructions on how to use it. You only do this once.

Before you paste
  • Start a new session in the Image Studio environment.
  • Claude asks for permission before it runs anything, so click Allow.
The prompt // paste into a new session in image studio
Set up this folder as my image studio, so I can create and edit images just by asking.

My Google API key is saved in the GEMINI_API_KEY environment variable and my OpenAI API key is in OPENAI_API_KEY. Never print the keys, and never save them in a file.

1. Build a small image tool in this folder that can:
   - Create images with Google's Nano Banana Pro (the default), Google's Nano Banana 2.1 (when I ask for a cheaper or draft image), and OpenAI's GPT Image 2.5 (when I ask for GPT Image).
   - Edit images: take one or more existing images as references and change only what I describe.
   - Set the shape (square, landscape, portrait or any aspect ratio) and the size (1K, 2K or 4K).
   - Save every image in an "images" folder with a short descriptive file name, without ever overwriting an older image.
   - Keep a log in images/costs.csv with the date, file name, model, size and estimated cost of every image.

2. Write a CLAUDE.md file that explains how to use the tool, so every future chat in this folder knows what to do: when I describe an image, turn my description into a detailed prompt, run the tool, and show me the result. When I ask for a change, edit the latest image unless I name a different one.

3. Test both keys with one small, low-cost image each, and show me the results.
What you should see

One test image from each key, so you know everything works. In the video, it made a banana on a wooden table with Google's key and a red paper boat with OpenAI's key. Both test images together cost about 4 cents. From now on, any chat you start in this environment can make images.

Step 5

Create Your First Image

You don't need any special wording. Describe it the way you'd brief a photographer: what's in the shot, where it is, the light, and the mood. Claude turns your description into a detailed prompt and sends it to Nano Banana Pro.

The prompt // new session in image studio
Create a square photo for Instagram. A small coffee shop called Copper Cup is launching a new fall drink, the Maple Cinnamon Latte. Show the latte in a ceramic cup on a wooden table by a window, with a cinnamon stick and a few maple leaves next to it. Soft afternoon light, cozy fall mood, shallow depth of field, like a professional food photo.
What you should see

About 15 seconds later, a latte with leaf-shaped latte art on a wooden table, a cinnamon stick and maple leaves next to it, and afternoon sun through the window. The image is saved in the images folder, and if you click it in Claude's reply, it opens right next to the chat. It cost about 13 cents.

Step 6

Edit Any Image

If something's not quite right, you don't have to start over. Tell Claude what to change, and Nano Banana keeps the rest of the picture the same.

1. Change one thing // same chat
Change the cup to a copper-colored ceramic mug with "Copper Cup" printed on it in small white letters. Keep everything else the same.
  • In the video, the cup turned shiny copper with the name in clean white letters, and the latte art, leaves, cinnamon stick and light stayed exactly where they were.
2. Edit your own photo // drag a photo into the chat first
  • Drag a photo into the chat, or drop it into the Image Studio folder. You can change the background, the lighting or even the season.
Here's a photo of me. Put me behind the counter of a cozy coffee shop, wearing a dark green apron and holding a cup of coffee. Keep my face exactly the same. Make it a landscape photo.
What you should see

You, with the same face and hair, in front of an espresso machine and shelves of coffee bags. Each edit cost about 13 cents.

Step 7

Add Text With GPT Image

Just ask for GPT Image in your message, and Claude switches to your OpenAI key. GPT Image is good at text, so it's the one to use for a poster.

The prompt
Use GPT Image to make a portrait poster for the counter of Copper Cup. Headline: "New for Fall". Below it: "Maple Cinnamon Latte" and the price "$5.75". Warm fall colors, a latte with maple leaves, clean and modern design.
What you should see

A clean poster with "New for Fall" at the top, a latte with maple leaves in the middle, and "Maple Cinnamon Latte" and "$5.75" underneath, every word spelled right. It cost less than a penny.

Which model to use
  • Nano Banana Pro: the best photos and the most reliable edits.
  • Nano Banana 2.1: great for quick drafts, at about a quarter of the price.
  • GPT Image: when the picture needs a lot of text.
  • Mix the two: ask Nano Banana for the photo, then ask Claude to put the headline and price on top in code. You get a realistic photo with text that's always perfect and that you can change in seconds.
Step 8

Build A Website With AI Images

Because everything happens inside Claude, your images don't have to stop at single pictures. They can go straight into bigger projects.

The prompt
Build a simple one-page website for Copper Cup, a cozy neighborhood coffee shop. Create a wide hero image for the top of the page with Nano Banana Pro, showing the inside of the shop in the morning, and use it on the page. Keep the design warm and simple.
What you should see

In the video, Claude wrote a prompt for a wide photo of the shop in the morning sun and kept the left side darker on purpose so the headline would be easy to read. Then it built the page around it and used the copper mug photo in the menu section. The hero image cost about 13 cents.

More ideas
  • Featured images for blog posts.
  • Product photos on clean backgrounds for an online store.
  • Ads in every size Facebook and Instagram ask for.
  • YouTube thumbnails.
  • To see what you've spent, ask Claude how much your images cost this week. The tool logs every image and its cost in images/costs.csv.
Costs

What Each Image Costs

You pay the model makers directly, per image. In a month where you don't make anything, you don't pay anything.

Price per image // direct from google and openai
Nano Banana ProAbout 13 cents, or 24 cents in 4K
Nano Banana 2.1About 3 cents
GPT Image 2.5About 1 cent at medium quality
  • 50 images a month with Nano Banana Pro is less than $7. With the cheaper models, it's closer to $2.
  • Everything made in the video, including the two test images, cost 59 cents.
Other pay-as-you-go options
  • kie.ai sells the same models with no subscription. When the video was made, Nano Banana Pro cost about 9 cents there.
  • Open Higgsfield is Higgsfield's developer platform, where you pay per image, or per second of video, instead of buying a plan.
  • Claude can work with any of them, because they all give you an API key. The setup works the same way.
Stay Safe

Look After Your Keys

  • Treat your API keys like passwords. Anyone who has them can spend your money.
  • Never paste a key into a chat. Put it only in the environment's Environment variables box, where Claude keeps it encrypted on your computer.
  • Set a spend cap. Google's monthly spend cap and a small prepaid OpenAI balance mean you can never spend more than you planned.
  • The setup prompt tells Claude never to print the keys and never to save them in a file.

Get the full system for building with AI, including the workflows, the prompt library, and everything that would not fit inside one video.

AI Builders Lab →